I'm an undergraduate student and I'm programming some HumanSeg iPhone app. Now I have to read raw frames from the camera, but I found the codes in official guide aren't clear enough for me to understand.
When I get the frames (in CMPixelBuffer format) from the camera, I need to modify it (I mean I have to do some padding and resizing, and turn it into a CVPixelBuffer format, to feed it to a CoreML MobileNet).
I've been searching for the solution for weeks, but unfortunately I got nothing. In the offical guide I was told that these buffers "doen't offer direct access to inner data".
I even tried to use a Context to draw the CMPixelBuffer into a UIImage and draw it back to a CVPixelBuffer，and I found this process terribly slow, just as the offcial guide says. Since I'm doing video processing this method is unacceptable.
What am I supposed to do or read? I really appreciate your help.
